The journal PLoS Pathogens yesterday retracted the first ever paper about XMRV, the virus that has become notorious for its reported—and now firmly refuted—link to chronic fatigue syndrome (CFS). The retraction, issued by the journal's editors, surprised the authors of the 2006 study, which reported a different link between XMRV and prostate cancer. The authors didn't think the paper needed to be retracted and weren't consulted by the journal or informed that the retraction would happen.
The retraction follows the publication of a new study yesterday in PLoS ONE, in which the authors of the retracted paper—along with others—conclude that the 2006 results don't hold up and report on an in-depth investigation on what went wrong in their labs. "But the discovery of XMRV, a new virus, still stands," says one of the authors, Robert Silverman of the Cleveland Clinic in Ohio, and an erratum to withdraw the prostate cancer link would have sufficed. "Why retract results that are valid?" Silverman asks.
The authors have already paid a price for their decision not to retract the paper voluntarily, according to Charles Chiu of the University of California, San Francisco (UCSF), the senior author of the new study. Several other journals with a higher impact factor than PLoS ONE had agreed to publish the paper, but on the condition that the authors retract the 2006 study—which the team refused unanimously, says Chiu, who wasn't involved in that study himself.
The alleged link between XMRV and prostate cancer was one of a few remaining loose ends in the XMRV story. Yesterday, a large multicenter study published in mBIO confirmed, again, that there is no link between XMRV, or a group of related viruses, and CFS. The paper to first report that association, published in 2009 in Science by Judy Mikovits of the Whittemore Peterson Institute for Neuro-Immune Disease in Reno, Nevada, and colleagues, had already been retracted, as had a similar 2010 paper in the Proceedings of the National Academy of Sciences.
Silverman's work on XMRV had started much earlier, around 2004. His group suspected that a virus might be involved in some cases of prostate cancer because men with mutations in the gene for RNase L, which is involved in innate immunity, have a higher risk for the disease. A collaboration with UCSF researchers Joseph DeRisi and Donald Ganem, who had developed a virus discovery microarray called the ViroChip, then turned up XMRV.
When the CFS study unraveled, Silverman says he realized that his results, too, might be in trouble. A series of papers published in 2010 and 2011—most notably, a study by Vinay Pathak and John Coffin of the National Cancer Institute in Frederick, Maryland—convinced him that the study was wrong and that XMRV had accidentally infected samples in the lab. From then on, "I felt I couldn't rest until I figured out how it happened," he says. "I wanted to get some closure."
For the new study, Silverman teamed up with Chiu and John Hackett of Abbott Laboratories in Abbott Park, Illinois. The team took tumor samples from 39 new prostate cancer patients and tested them for XMRV using three different techniques; they also went back to tumor tissue still available from the 2006 study. This time, they found no XMRV in any of the samples. They did find it in archived RNA extracts from the 2006 study, indicating that contamination had happened during sample processing.
Further studies—some using techniques unavailable at the time of the study—revealed that the virus originated in LNCaP, a cell line infected with XMRV that Silverman's lab used for other studies. The LNCaP cells, in turn, had become contaminated by 22Rv1, another widely used cell line that is now known to harbor XMRV.
The new study wins praise for its meticulous effort to set the record straight. "This paper made me feel proud of the authors and of our profession," Pathak says. "These scientists put their egos aside and aggressively and relentlessly pursued several lines of investigation to get to the truth." The team "deserves a medal," adds Kim McCleary, head of the CFIDS Association of America, an advocacy group for CFS patients. In the long history of pathogens falsely blamed for CFS, McCleary says she's never seen scientists retrace their steps so scrupulously.
In the paper, the group writes that XMRV is still a "genuine infectious agent" with "as-yet undefined pathogenic potential"; they point out that the virus is able to infect mice and two primate species and that several research groups have discovered interesting biological properties that could be useful, for instance, in cancer research. That's why an erratum to the original study would have been enough, Silverman says. "We've corrected the literature," he says.
Silverman notes there are precedents; in 1997 for instance, a research group reported finding a new human retrovirus, HRV-5; when that turned out to be a rabbit virus, they wrote a new paper to set the record straight without retracting the old one. Chiu adds it's odd for future papers on XMRV to refer to a retracted paper as a description of the discovery of the virus.
Silverman says the journal didn't contact him to discuss the retraction. DeRisi and Ganem, the two other senior authors on the paper, weren't contacted either, says Chiu, who says he has communicated with them. (DeRisi and Ganem did not respond to questions from ScienceNOW.) "I have never heard of a unilateral retraction like this," Chiu says. "They should at least have the courtesy of informing the authors in advance." Chiu says he was in the process of crafting an erratum with the authors.
"This is an unusual case," confirms Arturo Casadevall, a microbiologist at Albert Einstein College of Medicine in New York City and the editor-in-chief of mBIO. "Usually the authors are involved in the decision and on most retractions at least some of the authors agree with the retraction." Guidelines from the Committee on Publication Ethics stipulate that for retractions, "[w]henever possible, editors should negotiate with authors and attempt to agree a form of wording that is clear and informative to readers and acceptable to all parties."
A spokesperson for PLoS Pathogens couldn't answer questions about the retraction this morning but said the journal would respond later today or tomorrow. The retraction acknowledges that "the original finding of a novel gammaretrovirus, XMRV, with the use of a pan-viral detection microarray is valid, and sequencing and phylogenetic characterization of the virus still stands." It doesn't explain why the entire paper was nonetheless retracted.
The publication of the new paper put Chiu, the senior author, in an odd position, because he wasn't involved in the 2006 study or the debate on whether to retract it. Chiu says he submitted the new paper to two other journals before going to PLoS ONE. "It was a bit ridiculous," he says. "The reviewers all said, 'Great, this needs to be published.' But the editors said what prevented them from publishing was that the authors refused to retract the original paper." Chiu declined to name those journals.
In response to questions, Cory Mann, a spokesperson for PLoS Pathogens, sent ScienceNOW an e-mail that contained the following statement:
"The corresponding author was contacted by email on August 27th regarding our decision, and was asked to comment on the retraction text and suggest changes. We did not receive a reply, and decided to move forward with the retraction in conjunction with the PLOS ONE publication yesterday, September 18th."
The retracted paper had two corresponding authors, however: Silverman and DeRisi. Neither Mann nor the journal's editor-in-chief, Kasturi Haldar, responded to multiple follow-up e-mails. Silverman says he did not receive the 27 August e-mail or any other e-mail to the same effect; DeRisi did not respond to multiple requests for comment.
*11:06 a.m., 21 September: This item has been updated to include a statement from Cory Mann, a PLoS Pathogens spokesperson.